This paper discusses the incorporation of heuristics in computerized decision aids that are used for structuring decision problems. Decision theory is described in terms of linked subsystems, and sites where it is necessary to employ heuristic rather than normative, algorithmic procedures are identified. Production systems are then used to illustrate heuristic devices handling transactions across the interface between decision-theoretic subsystems and the decision maker's semantic memory. It is suggested that these heuristics can be incorporated into decision aids designed to improve the quality of access to the information contained within the decision maker's semantic memory. Candidate heuristics for use in development of such decision aids are identified and criteria for locating the operation site of heuristics are developed. The preliminary implementation of heuristic aiding devices within one particular decision aid, Multiattribute Utility Decomposition (MAUD), is briefly described. Heuristics that operate on the interface with semantic memory are identified within a comprehensive table of heuristics cited in the literature. Also provided are guidelines for further development of automated decision structuring systems by the use of modular uncertainty structures, and a 98-item bibliography. (Author/ESR)